What working with us costs.
Fixed prices per milestone or monthly retainers — never an open meter. The bands below are where engagements typically land, and a scoped quote gets you the exact number before you commit to anything.
We price the way we build: fixed outcomes, no meters running. Every engagement is priced per milestone — scope agreed in writing, price fixed before work starts — or as a monthly retainer with defined outputs. You will never receive an hourly bill from us, because hours are our problem, not yours: agents absorb the mechanical work at near-zero marginal cost, and the senior hours that remain are priced into the milestone, not counted against it.
The bands below are where our engagements typically land. They are ranges, not quotes — your exact number arrives as a written, scoped proposal after one 30-minute call, and it is the number we hold ourselves to. If your problem does not fit our model, we will say so on that call rather than shape it into one that does.
Engagement models and price bands
Assessments & audits
AI risk & security assessment · workflow audits
- — AI inventory: sanctioned and shadow systems, mapped with data flows
- — Risk register scored by likelihood and impact
- — Prioritized remediation roadmap
- — Policy pack your legal team can adopt
2–4 weeks typical; a single-team first-pass inventory takes days.
Explore AI risk & security →Build engagements
software development · AI workflow automation
- — Shipped, monitored software or automated workflows — not a repository of promises
- — Tests and CI, generated by agents and enforced by humans
- — Handover documentation your team can extend without us
- — A named senior engineer accountable kickoff to launch
3–6 weeks to a first production release.
Explore build services →Team upskilling
engineering and operations tracks
- — Curriculum built on your stack and your team’s real work
- — Every participant ships a real automation, supervised
- — Written playbooks your team keeps
- — Before/after measurement of adoption and time saved
4–6 weeks per cohort; 2-week intensives for deadline pressure.
Explore AI upskilling →Agent Ship Week
one agent shipped to production · fixed price
- — Day-0 opportunity mapping from your docs, so your week starts with correction, not generation
- — Constraints, human gates, audit trail, and rollback encoded as system enforcement
- — Production go-live on day five — you keep the agent, the case book, and the measurement loop
Five days to production go-live, plus a 30-day measured check-in.
Explore Agent Ship Week →All figures in USD, before applicable taxes — except Agent Ship Week, which is a fixed £5,000. First milestones can land below these bands; full programs can run above them. Bands are honest medians of where engagements land — the scoped quote is the number that matters.
What moves the number
Drives it up
- Integrations
- Every additional system in scope — ERP, CRM, ticketing, internal services — adds wiring, testing, and failure modes.
- Regulated data
- Compliance regimes (BAA scope, EU data residency, audit trails) add design and review work to every milestone.
- Deadline compression
- Parallel work costs more than sequenced work, in any delivery model. Calendars you control are cheaper than calendars imposed.
- Legacy archaeology
- Undocumented codebases need reconstruction before they can be extended safely — the archaeology is real work and priced as such.
- Cohort size and tracks
- Upskilling scales with people: bigger cohorts and parallel engineering/operations tracks move the number up.
Keeps it down
- A narrow first milestone
- One workflow automated or one production slice shipped keeps the first number small — and lets you judge the model on your ground.
- Clean APIs
- Where your systems already expose interfaces, integration stops being the budget sink it usually is.
- Fast feedback
- A reviewer on your side who answers within a day keeps every delivery cycle short. Slow approvals are the quiet cost multiplier.
- Standard stacks
- Boring, well-documented technology is cheaper for agents and humans alike; bespoke infrastructure is where budgets go to hide.
- Shared cohorts
- Training two teams together in one cohort costs meaningfully less than two separate programs.
The arithmetic behind these bands, in full: our cost model, how we compare to agencies and freelancers, and the savings calculator for the payback math against a build band.
How you get an exact number
A 30-minute call
The problem, the calendar, the constraints. A working conversation, not a pitch — and if we are the wrong shape for the problem, we say so here.
A scoped proposal
Scope, milestones, and a fixed price for each — in writing, before any work starts. That document is the number we hold ourselves to.
A first milestone
Small on purpose: one workflow automated, one slice shipped, one assessment delivered. Evidence beats proposals.
The shape that fits
Next milestones or a monthly retainer — your call after the first one, with no obligation either way.
Pricing questions we get asked
Why do you publish bands instead of exact prices?
+
Because scope moves the number more than the delivery model does — an assessment of one team is a different universe from an assessment of a multinational. The band tells you whether we are in the same budget conversation; the scoped proposal gives you the exact figure, in writing, before you commit to anything.
Do you bill by the hour?
+
No. Every engagement is a fixed price per milestone or a monthly retainer with agreed outputs. Hours are an internal concern — and a misleading one, because agents do most of the mechanical work at near-zero marginal cost. Metering hours would mean charging you more for our automation, not less.
Are AI model and tooling costs included in the price?
+
Our fee is fixed and covers our work. Third-party costs — model API usage, cloud infrastructure, licensed tools — are estimated and itemized in the proposal so you always see them apart from our fee. In practice they are a small share of a typical engagement.
What does the first call commit me to?
+
Thirty minutes and nothing else. It is a working conversation about your problem, calendar, and constraints — not a sales sequence. If we are not the right fit, telling you that on the first call is the cheapest outcome for both of us.
Do you offer retainers?
+
Yes. Continuous work — extending an automation program, ongoing delivery, post-launch support — runs on a monthly retainer with outputs agreed in advance. Most retainers start after a first milestone has proven the model on your ground.
Why not an agency, or fully autonomous AI?
+
An agency funds a full pyramid, including the layers that coordinate it. A fleet of autonomous agents is nearly free per hour and dangerous per outcome — in our assessment work we regularly meet teams that saved heavily on generation and then spent multiples of it on remediation. Our comparison page and cost model walk through where each model wins.
Want your number, not a band?
The bands are honest medians. Nobody should budget against a median — one call gets you a written, scoped price.